Practical Predictive Analytics

An example – airquality

In this example, we will use the regsubsets() function contained within the leaps package to determine which variables are important for predicting temperature:

install.packages("leaps") library(leaps) data(airquality) str(airquality)  

In the following regsubsets() function call, we specify that we want the best model for each combination of one, two, and three variables. By default, the function will compute the best model for all variables in the model. In the function call, I illustrate specifying a max value, since I want to be able to stop the algorithm after a certain number of variables has been reached. This becomes important when you use data that has a large number of variables, and you do not want to tie ...